<robmur01>
sadly I'm not sure I can even find the original Windows CE drivers for the SDIO card any more :(
kevery has quit [Ping timeout: 268 seconds]
kevery has joined #linux-rockchip
kevery has quit [Client Quit]
<CounterPillow>
perhaps they are better left lost to time
<macromorgan>
alyssa: I figured out the issue with U-Boot being persnickety when it comes to holding the reset button; if you want to try a New and Improved Image™ I can make one now.
<macromorgan>
long story short... if you don't want to boot off the serial flash controller, don't include the driver for the serial flash controller. I'm sure there's a better fix in there somewhere but for now that seems to work consistently/reliably no matter how long I hold down reset.
<alyssa>
marvs: Weeeee
<alyssa>
I mean macromorgan
<alyssa>
macromorgan: meanwhile my CTS run failed after 11hrs because apparently 1GB of RAM isn't very much
<macromorgan>
yeah, it's not
<macromorgan>
I use zram and swap proper (on an SD card, yuck) to try and stretch it as far as I can get it to go.
<alyssa>
good time to fix panfrost's memory allocation behaviour to not just `assert(malloc() != NULL)` lol
<macromorgan>
also not sure if you heard me but those patches you sent yesterday tested fine/fixed my issue
<macromorgan>
also figured out my issue with getting phosh running. Which I eventually got running and let's just say yeah, you guys were right.
<alyssa>
macromorgan: awesome
<alyssa>
and not awesome
<alyssa>
respectively
<macromorgan>
I think ludo could be a great inferface, if only 1) it supported aarch64 (I assume I can "make" it support aarch64 if I really want) and 2) it supported launching external programs like a media player/tux racer/doom/etc.
hipboi has joined #linux-rockchip
<macromorgan>
I just learned C this year though, and don't feel like adding Go onto the list of things I know quite yet.
<macromorgan>
would a new image help you at all? If so I can take the time to put one together and test it, otherwise I won't bother.
<macromorgan>
still doesn't support video or optee, but at least the reset button isn't wonky anymore
<macromorgan>
(note that video works fine once linux proper takes over)
<alyssa>
this image is working well :-)
<macromorgan>
okay cool, if it ain't broke focus on the things that are...
<macromorgan>
I guess for my next trick I get to figure out how to get upstream U-boot's SPL to boot upstream Optee.
hipboi has quit [Ping timeout: 248 seconds]
<CounterPillow>
I'm currently debating whether I V2 my patchset today or whether I'll wait for comments on the actual C code.
hipboi has joined #linux-rockchip
<robmur01>
FWIW, having found a PX30 TRM out of curiosity, I'm starting to wonder whether the 2-channel version in particular is all that different from what's in previous SoCs, and whether it might make more sense to essentially just add the 8-channel TDM support to the existing driver
<CounterPillow>
Time for me to compare my RK356x TRM to some other TRM that has I2S I guess.
hipboi has quit [Ping timeout: 250 seconds]
<CounterPillow>
TXCR register looks compatible, all the additional bits they used were reserved previously.
<CounterPillow>
on the other hand, I'd have to rewrite an entire working driver and risk breaking another working driver in the process.
<CounterPillow>
I'll ponder it, will need to make sure the registers aren't too wildly different and then look into how compatible the mainline I2S code is to being extended like that
<CounterPillow>
Nice of them to have a VERSION register but make it at a different location for the two versions
<alyssa>
are the rockchip TRMs public these days?
<CounterPillow>
Only some. The one I have isn't public, but the RK3328 TRM I'm comparing to is public (well at least the first part of it is)
<alyssa>
Ah, yes.. I seem to recall the RK3399 TRM having both public and private versions with the public one missing most of the interesting stuff :V
<CounterPillow>
I wish vendors would just publish all their documentation, it's not like they're hiding anything secret from everyone else making almost identical ARM SoCs
<alyssa>
Seriously.
<CounterPillow>
I'm glad Rockchip is at least dipping their toes in it though, it's a step.
<CounterPillow>
Though if I was authoring an 850 page document in MS Word I'd probably also be too ashamed to publish it c:
<alyssa>
hehehe
<alyssa>
I've only written hw docs once it was in XML I mean Markdown I mean LaTeX
<alyssa>
#shameless
hipboi has joined #linux-rockchip
<CounterPillow>
Okay here's a wild idea: I make my i2s-tdm driver the new i2s driver because from looking at the code it's already better than the existing i2s driver, which has a lot of the stuff I refactored in mine
<robmur01>
AFAIK the usual issue is that the 3rd-party IP vendors get sniffy about having *their* documentation republished in a public manual, and most SoC vendors can't be bothered negotiating that